Framework for Analysis

Issued February 26, 2009

Executive Summary

1. This report has been prepared in the context of the proposal by the European Commission to introduce functional separation as an “exceptional remedy” and in the light of proposals in various other countries, for example New Zealand and Australia.

2. There has been considerable debate about the merits or otherwise of functional separation, but much of it does not put functional separation in its proper context and fails to analyse the role it plays in an overall regulatory environment designed to address competition problems.

This report seeks to answer three key questions:

i) What is it about the structure of electronic communications markets that gives rise to enduring competition concerns?

ii) Have equivalence and functional separation been well-designed and implemented in the UK to address enduring competition problems and what does this imply for other fixed communications markets?

iii) Can we expect equivalence and functional separation to lead to improved intermediate (wholesale) and final consumer outcomes?

3. This report seeks to answer three key questions:

i) What is it about the structure of electronic communications markets that gives rise to enduring competition concerns?

ii) Have equivalence and functional separation been well-designed and implemented in the UK to address enduring competition problems and what does this imply for other fixed communications markets?

iii) Can we expect equivalence and functional separation to lead to improved intermediate (wholesale) and final consumer outcomes?

4. Before the liberalisation of electronic communications markets starting in the UK in 1984 and implemented throughout Europe in 1997 (as well as being adopted in other countries), the market was characterised by vertically-integrated monopolies. In Europe, National Regulatory Authorities (NRA) were granted powers to impose ex ante regulation on operators with Significant Market Power (SMP) in relevant markets, to prevent them from discriminating against their competitors when those competitors were buying essential inputs from the SMP operator.

5. When Ofcom conducted its Telecoms Strategic Review (TSR) in 2004 it summarised its findings by saying that those who rely on BT to provide access have experienced twenty years of:

  • slow product development;
  • inferior quality wholesale products;
  • poor transactional processes; and
  • a general lack of transparency.

6. Despite there being no finding against BT on non-discrimination grounds, Ofcom determined that non-price discrimination was the major problem facing the market. It also found that a reliance on an obligation of non-discrimination (which anyway allowed discrimination when objectively justified) and accounting separation was not sufficient to support a competitive downstream market and the consumer benefits that would flow from this.

7. Ofcom’s preferred remedy was to ensure that BT offered “real equality of access”, such that both internal and external downstream customers of upstream essential facilities were provided with the same product, on the same terms and using the same ordering system. Ofcom also wanted to see a re-organisation of BT to support the delivery of equality of access.

8. In lieu of a referral to the Competition Commission under the Enterprise Act, BT offered, and Ofcom accepted, a set of Undertakings whereby BT would offer Equivalence of Input for a defined set of products and would create a separate Access Services Division (later branded Openreach). “Functional Separation”, as this organisational structure became known, was introduced to support equivalence.

9. We have conducted a range of interviews with organisations involved in the telecoms market in preparing this report. There was general agreement amongst the interviewees that whilst the implementation of equivalence and functional separation has not been perfect, it has brought about considerable improvements. Most importantly, downstream competitors of BT now have more confidence to invest, and indeed are doing so.

10. Drawing on the experience in the UK, we set out a qualitative framework for the assessment of the effectiveness of equivalence and functional separation in which we examine the impact of these measures on:

i) Investment and innovation – which can be broken down into two areas: in the local loop; and in downstream markets by all operators;

ii) The internal efficiency of the regulated firm; and

iii) The direct financial costs of regulation.

11. A number of authors have claimed that equivalence and functional separation will negatively affect investment because, for example, the upstream firm will not be able to earn a sufficient return and that a downstream firm will behave strategically if and when investment is made: this is often referred to as the “hold-up problem”.

12. We find these arguments to be misplaced. First, a firm with SMP or dominance in an upstream market is almost certainly going to be subject to price control regulation, regardless of whether it is also subject to equivalence and functional separation. If price regulation does affect investment decisions, this is a general argument against regulation rather than against any particular form of regulation.

13. Secondly, for the hold-up problem to occur, the investment made by the upstream firm has to be specific to one downstream firm. This is simply not the case in electronic communications, where there are multiple firms competing downstream and so multiple outlets for an upstream product.

14. Looking at what has happened in the UK, we find that in 2008 both BT and Virgin Media made substantial announcements about their investment plans for the local loop. BT is to invest £1,500 million to deliver fibre to 12 million homes (mostly fibre to the cabinet), whilst Virgin Media is upgrading its HFC network to DOCSIS3, which will deliver 50 mbit/s by the end of 2009. We also see that broadband Internet Service Providers have invested in unbundling local exchanges and installing their own DSLAMs to offer higher access speeds. The UK has five million LLU lines today, compared with fewer than 200,000 when BT and Ofcom signed the Undertakings.

15. The internal efficiency of the regulated firm may be negatively affected by the specific implementation of functional separation. We found examples of efficiency losses and gains in BT. At a systemic level though, we find no reason why functional separation should lead to an efficiency loss, in particular because the firm remains integrated. It is therefore able to realise the efficiency gains of integration whilst ensuring equivalent treatment to its wholesale customers.

16. With regard to the direct cost of regulation, if equivalence delivers competition benefits downstream, then such markets should cease to be subject to SMP and so ex ante regulation can be withdrawn. We have already seen this in the UK, where the growth in LLU has allowed Ofcom to withdraw regulation in the Wholesale Broadband Access Market in about 65% of the country.

17. Given that functional separation as a remedy is being considered in many countries, we have set out a framework for assessing whether equivalence and functional separation are necessary and for determining whether the proposed form of functional separation is likely to deliver benefits in both intermediate and final consumer markets. We have tested this framework using the proposals in Australia and Italy.

18. Finally we return to the three questions set out above. We find that:

i) The economics of the local loop mean that, at least in the medium term, there are likely to be enduring competition problems at the access level, which will need to be regulated to ensure competition downstream.

ii) The implementation of equivalence and functional separation in the UK has been largely successful in addressing enduring competition problems in local access markets that serve residential consumers. However, there was less satisfaction expressed with the level of attention paid to wholesale products for business customers. This probably represents priorities agreed with the regulator and hence a limitation of the implementation of, rather than any fundamental problem with, equivalence and functional separation.

iii) Equivalence and functional separation are not cost-free remedies and may impose marginal costs on the regulated operator that it would otherwise not incur. However, the arguments put forward that these remedies will damage investment incentives appear to us to be erroneous. We conclude that the removal of discrimination leads to greater dynamic efficiency gains, which outweigh static efficiency losses to the extent that they occur.

The full discussion document can be downloaded as a PDF document PDF document (369kb)